How to change the dimension of a numpy array-CodePudding

I have an array of 2-d values created through a mesh grid. The arrays are returned as 40X40 objects that I want to transform to 2X1 array in order to perform multiplication with 2X2 matrix.

A=np.array([[0,-1],[1,0]])
x, y = meshgrid(arange(-2, 2, 0.1), arange(-2, 2, 0.1))
X=np.array([[x], [y]])

np.matmul(A,X)

I want the multiplication to return a 2X1 vector with the first component being all the -y values and the second being x values.

EDIT: I want to point out that the matrix multiplication is not absolutely necessary. The point is that I want to dictate whether I get y or x through a matrix which also dictates the sign of the values and their magnitude.

CodePudding user response：

Your array's shape will be 40x1 not 40x40. It cannot be converted into 2x1. But it can converted into the multiples of 40 shape. Like 8x5,20x2,10x4 etc.

CodePudding user response：

You can reshape the 3D 2x40x40 array to be a 2D 2x1600 array, and use matmul on that, reshaping the result back afterwards:

X = np.array((x, y)).reshape(2, 1600)
A = np.array([[0,-1],[1,0]])
Y = (A @ X).reshape(2, 40, 40)